use output of previous magrittr chains as arguments to further arguments

Question

if I have the following example:

library(text2vec)
library(magrittr)

reviews <- movie_review[1:10,]

vocabInsomnia <- reviews$review %>% itoken(tolower, word_tokenizer, n_chunks = 10) %>%
    create_vocabulary %>%
    prune_vocabulary(term_count_min = 10, doc_proportion_max = 0.5) %>%
    vocab_vectorizer %>%
    create_dtm(<output_from_itoken>,<output_from_vocab_vectorizer>)

You can see that in the very last chain sequence I want to use the outputs of two of the previous steps as arguments to the create_dtm function. I only know how to feed in the output from the chain directly before i.e. output from vocab_vectorizer, but not the output from the function itoken that was the first chain in the sequence. Does magrittr allow this?

score 5 · Accepted Answer · answered Jan 18 '22 at 17:01

We could create a temporary object using pipeR

library(text2vec)
library(pipeR)
library(magrittr)
reviews$review %>% 
  itoken(tolower, word_tokenizer, n_chunks = 10) %>>%
 (~ tmp) %>%
  create_vocabulary %>% 
  prune_vocabulary(term_count_min = 10, doc_proportion_max = 0.5) %>% 
  vocab_vectorizer %>% 
  create_dtm(tmp, .)

-output

10 x 6 sparse Matrix of class "dgCMatrix"
   an so by are he br
1   2  4  .   2  9  8
2   .  1  1   .  .  .
3   1  .  6   7  .  2
4   4  1  3   2  .  4
5   2  .  1   1  .  .
6   .  .  .   .  .  .
7   1  3  .   .  .  .
8   .  1  .   .  2  .
9   .  .  .   .  1  4
10  .  .  .   .  .  2

Nice! This is very helpful and will immediately become a staple of my workflows! — Greg, Jan 18 '22 at 17:02

score 3 · Answer 2 · answered Jan 18 '22 at 16:51

I don't know if there's a cleaner or more efficient way to do this, but what I usually do in this situation is to nest piplines at the highest level where I need to pull an input from and pipe in the output using . to continue the chain.

library(text2vec)
library(magrittr)

reviews <- movie_review[1:10,]

vocabInsomnia <- reviews$review %>% 
  itoken(tolower, word_tokenizer, n_chunks = 10) %>%
  create_dtm(., create_vocabulary(.) %>%
               prune_vocabulary(term_count_min = 10, doc_proportion_max = 0.5) %>%
               vocab_vectorizer())
  
vocabInsomnia
#> 10 x 6 sparse Matrix of class "dgCMatrix"
#>    an so by are he br
#> 1   2  4  .   2  9  8
#> 2   .  1  1   .  .  .
#> 3   1  .  6   7  .  2
#> 4   4  1  3   2  .  4
#> 5   2  .  1   1  .  .
#> 6   .  .  .   .  .  .
#> 7   1  3  .   .  .  .
#> 8   .  1  .   .  2  .
#> 9   .  .  .   .  1  4
#> 10  .  .  .   .  .  2

^{Created on 2022-01-18 by the reprex package (v2.0.1)}

use output of previous magrittr chains as arguments to further arguments

2 Answers2